The PI is a quantitative researcher with a background in computer science and a record of publications in the areas of image analysis methodology, computer vision and machine learning. His objective is to establish himself as a highly productive independent researcher of Alzheimer's disease (AD) and dementia by complimenting his strengths in quantitative analysis with a newly acquired expertise in the biomedical aspects of dementia and aging. Prof. John Detre MD, the head of Penn's Center for Functional Neuroimaging and an internationally renowned neurologist, has offered to serve as the principal mentor. The co-sponsors, Prof. John Trojanowski, MD, PhD and Prof. Murray Grossman, MD, EdD have offered to help the PI acquire expertise in the pathology and physiology of dementia as well as its impact on memory, language and thinking. The research plan involves developing an advanced framework for detecting and tracking structural and functional changes in the anatomical subregions of the hippocampus and parahippocampal gyrus using in vivo neuroimaging. Using this framework, the PI will evaluate the hypotheses that AD-related changes differ across the anatomical subregions of these temporal lobe structures and, consequently, that their morphology and physiology can be used to predict AD early and accurately. These hypotheses will be tested by mining the massive database of longitudinal MRI image data of AD patients, people at risk for AD and elderly controls, which is being generated by the ADNI initiative launched recently by the NIA/NIH. RELEVANCE. As the baby boom generation ages, the number of US families devastated by AD and the associated financial burden on the society is expected to increase dramatically. While no cure for AD is known, a sizable effort is underway in developing pharmaceutical agents that may halt or slow down the neurodegenerative processes that cause AD. To be effective, these treatments will require early detection. The Pi's career objectives, addressed by the [unreadable] [unreadable] [unreadable] [unreadable] research proposed in this application, are to improve the accuracy of early AD diagnosis and to develop non-invasive analytic tools that would aid drug development and broaden what we know about the pathology and physiology of AD. [unreadable] [unreadable] [unreadable]